MIcompany launches AI leadership program for women

Dorthe van Waarden nominated for
Dutch Young Talent Award 2019

Over the past three years MIcompany has invested in strengthening her female employee group through a special leadership program. This internal program has met and exceeded expectations, resulting in a growth acceleration of the female leaders within MIcompany. Additionally, Laura Brandwacht – principal of MIcompany – was one of the finalists of the Young Talent Award 2018. This year, a second principal of MIcompany – Dorthe van Waarden – is nominated for this award. These and additional successes lead to our conviction that using AI to accelerate female leadership is an unbeatable combination. Therefore, MIcompany has decided to launch a new certified AI leadership program for women together with The Inclusion Factory in order to open up this success formula for other organizations and women as well.

The AI leadership program for women is a certified program of 5 badges

The success of our internal program and the growing demand for more female leadership within companies has led us to believe that opening up our AI leadership program for women might just create a real win-win situation. For both ambitious data driven employers and their talented and ambitious female professionals. The program enables women to rise to leadership roles using AI as a natural lever that plays directly into their strengths. It strengthens women with substantive and strategic Data & AI knowledge (a burning platform in many boards) as a foundation for their leadership. Additionally, it empowers participants to live up to their full potential by creating awareness of gender differences, the rules of the game and how to leverage their natural strengths with confidence and resilience.

The program is meant on the one hand for female managers (to be) with no prior AI background that want to strengthen both their AI leadership skills (because they are directly or indirectly involved in AI topics more and more) ánd who wish to learn how

to use their feminine strengths as a leader. On the other hand, women who have just started leading a data or analytics team who want to strengthen both their AI ánd their leadership skills (with or without an AI background) will in our experience also benefit greatly from the program.

The AI leadership program for women consists of 5 certified badges. Each badge covers AI leadership topics in combination with related gender intelligence subjects. Participants learn how women can utilize their feminine strengths in order to succeed as a leader, using AI to make an impact. And we will practice with real-life situations that women typically encounter in their work.

The program starts with Part 1, containing 2 badges: AI Foundation for Women and Leading in Diverse Teams. Part 2 of the program, aimed at participants who also aim to be involved in use case implementation consists of 3 more badges: Implementing AI solutions with Resilience, Building Long Lasting Relationships, and the choice of either the Ethics in Modelling or Data Strategy badge.

See the figure 1 for more details on the content of the different badges.

MIcompany and The Inclusion Factory are very excited to join forces and launch this first edition of the AI leadership program for women in the spring of 2020. Would you like to be a part of the program or do you know someone for whom this would be a perfect fit? Feel free to contact one of us (see next page) for more information.

Dorthe van Waarden nominated for Young Talent of the Year

Dorthe van Waarden (29) is nominated for the title Young Talent of the Year 2019 as part of the yearly Dutch top woman election (‘Topvrouw van het jaar’). Her nomination is the result of her strong growth as a leader within MIcompany in the exciting field of Artificial Intelligence. Dorthe is co-leader of the Advanced Analytics practice area of MIcompany in which she is responsible for the model management services portfolio. These services are aimed at developing and especially maintaining key AI models for customers such as Max (Israel), KPN and Dirk. This means that she is not only responsible for the quality and prediction power of algorithms, but also to prevent biases in these algorithms. Her tips on how to increase gender diversity with AI and how to prevent gender biases are included in this news item.

1. Use AI to detect (human) bias in data or decision making.

AI is designed to find patterns in data. This means that if this historical input data (think assessment scores) or human performance (think % that gets promoted) is biased, these biased patterns will be learned by AI. On the positive side, AI will record and provide transparency on explaining variables and performance scores.  Over time it will allow us to correct assessments and performance scores to build a system that creates true equality of opportunity. A second step that AI will bring us is, is that it will force us to make decisions on real data, in a much more transparent way. There are many tests that will help us to asses bias or even fairness of the selection procedures applied. Only through working with AI, we will fix corrupt systems that are built on human bias for assessment and performance, that otherwise will perpetuate for ever in wrong judgements through human bias and errors in capability assessments and performance evaluations.

2. Program AI to ignore demographic information of candidates.

When using an algorithm for recruitment purposes, specific choices need to be made to exclude information such as gender, race, age etc; factors that are typically related to human bias. The human mind is not able to ignore information that it is has access to, and humans are therefore prone to make decisions on discriminatory characteristics. AI has no problem to forget information, we just re-run the algorithm and filter out the undesirable information. However, sometimes only excluding these discriminatory factors in not enough, since demographic characteristics can be ‘hidden’ within other information, such as university or zip code. Therefore, also this kind of variables should be excluded if you want to create an unbiased algorithm.

3. Use ‘out of company’ data sources.

If bias in data (e.g. assessment or performance scores) cannot be easily fixed in a country or business area, there is a real opportunity to learn from other situations or countries that have been able to build a solid data set. One could argue that we should use the data from countries that are leading in female representation (such as Norway, Iceland or Israel), instead of using the data from countries that have many inhibitors for female success.

4. Build additional rules in the algorithm.

Another way to stimulate gender diversity within recruitment, could be to force gender equality with AI. For example, we could instruct the selection system to take ‘gender’ into account and give the assignment to select the best woman for the job.

However, such hard force might not be what we’re striving for. Still, giving specific instructions to the algorithm could work as a first step towards diversity. For example, by instructing the algorithm to give more weight to certain (female) capabilities, or by selecting the five best men and the five best women for a subsequent job interview. This could just be the additional support and encouragement that is needed to stimulate gender diversity.

5. AI can measure the attractiveness of work environments for men and women.

Although fairness is always important, the biased selection within recruitment is not the only issue. Another challenge lies within the available pool; female leadership within board representation is low, but also within the group of leaders that applies for these positions the ratio men/women is still quite skew. Similar situations can be seen within the field of software engineering and AI. Therefore, one of the other key issues that we need to fix (next to the fairness of selection) is the attraction of more female leaders on critical positions in society. Maybe the largest bias that should be attacked is that various educational programs, jobs and working environments are designed by men, and for men. And because of that they are less attractive to female talent. Therefore, we need to better understand how we can attack these hidden barriers for female breakthroughs.

The good news is that AI will bring us the solution over time, since with AI we can predict to which degree female leaders are attracted to different environments, and which factors are crucial. This will help us to change these environments and get more diverse leadership teams.